The last couple of versions of JMP have extended data visualization in so many ways and made it easier to create these graphics, too. One recent addition to Graph Builder is shapes. JMP is installed with a set of shape files for the geographic boundaries of world countries, states and counties in the USA as well as first-level divisions for Canada, China, the United Kingdom, France, Italy, Japan and Germany. You simply include a data column in which the levels are the standard name for a particular shape in a column of character data (e.g., “Maine” or “ME”), and JMP will recognize it when you put this column in the shape role of Graph Builder.
You can also create your own shapes. In this example, I want to visualize the quantitative response from each of 96 wells in a microtitre plate:
JMP does not have shapes for these wells. I had to make these shapes. You make shapes with a pair of JMP data tables: a Name data table and a XY data table. The name file looks like this:
Notice that the names that I will use in my analysis (e.g., “A1” or “D3”) are in the Well column. This name is associated with a particular Shape ID. For example, the name “A1” is linked to shape ID 1. This shape name definition is established by a column property:
The shape ID, in turn, is mapped to its constituent visual parts in the second file, shown here:
In this example, all of the shapes contain a single part. That is, each well is represented as a simple circle. Imagine, though, a case of a floor layout in a building. You might have many shapes for this case, and some of the shapes might require multiple parts for a proper representation.
Each part ID is associated with a set of X and Y coordinates. The set of coordinates represent the vertices of a polygon that will be drawn by Graph Builder to represent the part. I included 30 vertices for each part (well) so that it will look like a smooth circle. It would be a tedious job to create both shape files even with much less resolution. I wrote a script to automatically create both shape files.
You can store both shape files in the same folder with the data table to be analyzed, or you can place them where JMP searches for custom map shapes:
On Windows: C:\Users\<user name>\AppData\Local\SAS\JMP\Maps
On Mac: /Users/<user name>/Library/Application Support/JMP/Maps
Here is my data:
The response is in the Relative Potency column. The shape name is in the Location column. You can call this column anything you like. The magic happens because of the same column property as before, Map Role, but with a different setting that points to the first of the shape files:
Everything is set: The custom shape files were made and stored where JMP will find them, the shape names are included in my data set, and the column property links back to the name definition file. Here is my data in Graph Builder using these shapes:
It is easy now to relate the change in the response with the design of the experiment and the location of each sample or reaction. There are three large blocks and a gradual change within each block.
Would you like to use these microtitre well shapes? You can find them in the JMP File Exchange (download requires a free SAS profile). You can read more about using shapes in the JMP guide, Basic Analysis and Graphing, which is available through the Help system: select Help > Books.